Log in

View Full Version : madVR - high quality video renderer (GPU assisted)


Pages : 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 [822] 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329

Dodgexander
1st December 2016, 02:30
I see there's been a lot of discussion since I last posted about my issue but I have done some more testing, this time manually putting my PC into sleep. The results are different than when my monitor first goes to sleep and the PC sleeps afterwards.

With MadVR and zoom player or mpc-hc the results are the same.

If the PC does go to standby and return the video is always green and I have to restart the player for it to work again.

With mpc-hc using the EVR it doesn't happen.

It also doesn't happen 100% all the time I manually sleep but does when I leave my computer to sleep (monitor turned off first).

The bsod I get that reboots the PC before the PC is allowed to go into sleep points to ntoskrnl.exe 14a510 but when the PC bsods on return from sleep the bsod points to dxgmms2 blaming the nvidia driver.

For a while I didn't notice this was related to madVR so I've carried out various steps to solve it already. I've made sure there's no memory fault with my PC by running memtest86 for 48h, I've checked with prime 95 for 24h and I've monitored temperatures to make sure they aren't too high. First thing I did was do a complete clean install of the nvidia drivers.

I'm not sure why I have this silly problem but it would be great to get to the bottom of it.

Anyone have an idea what it could be? Is my graphics card faulty?

Sent from my Nexus 5 using Tapatalk

Sunspark
1st December 2016, 04:22
Catmull-rom is the same as Bicubic 50 which is soft. There is nothing wrong with using Bicubic 60 for chroma upscaling. I use it by choice with jinc for luma because it looks better than jinc/jinc to me, and also because when it comes to colour, catmull-rom and bicubic are the chromas that match the colours of nearest-neighbour. I don't know about the newest algos as I never tested and they require more gpu anyway. The goal is not to make stuff sharp as a tack. When image editing, good quality unsharp masking after downscaling stops at the point where the sharpen just barely becomes visible via preview on and off, this is usually on an case by case basis for dialing in the numbers so I do agree with the selection of 60 here, because 75 was already on the sharper side.

ryrynz
1st December 2016, 04:53
Anyone have an idea what it could be? Is my graphics card faulty?


I'd start with the system BIOS first. Reset it and test then look for an updated BIOS for the motherboard and GPU.

Backflash
1st December 2016, 08:52
If the PC does go to standby and return the video is always green and I have to restart the player for it to work again.


I had DP to miniDP and screen always bugged on me if I left player on pause and left pc to switch screen off. Screen didn't wake up 3 out of 10 times. I got DP-DP cable it went away(for now), so I think it's driver issue in my case. It is also happening only on w10.
Just do not leave player running when you need to step away from PC.
Also it's generally bad to let PC go to sleep, it somewhat works for laptops because battery life, but there is no need to do that with PC. So many things can go wrong with modern GPU drivers, probably that's what happening, you are not saving word document into RAM, it's a bit more complicated with rendering I think.
Set it do disable display and never go to sleep, most safe option.

StinDaWg
1st December 2016, 08:58
I've heard a lot of feedback about the new settings dialog organization. Thanks, guys.

But I don't remember anybody commenting yet on the new NGU-Low quality level. It's faster but softer than the old NGU-Low. But is it still "acceptable"? How's sharpness, aliasing, ringing etc for you? Is it worthy of the "NGU" sticker?
I think NGU-Low is great for 720p->1080p upscaling. Switching between NGU-Low and NGU-Very High I see little to no difference with 720p content, except for a very very (did I say very?) small increase in sharpness.

I am now able to run NGU-Low on 720p60 videos, where I couldn't use any NGU levels before with my HD 7770. It's also significantly faster than Jinc+SR2.

There is a bigger difference with SD videos. There is a noticeable quality increase between NGU-Low and NGU-Med, with then smaller increases all the way up to Very High. Surprisingly I am able to use NGU-High on 480p60 videos with render times around 10ms, so it's not much of an issue.

If you are somehow able to make NGU-Low sharper without increasing gpu usage, that would be good though.

Maybe Catmull should be used in place of bicubic150? I've never a/b them though. I thought we were all happy with Catmull, especially as as downscaler, for a long time :cool:
Please no. Bicubic 150 AR is a very fast and good alternative to SSIM. It's even better in some ways judging from the pics posted above (we are talking microscopic zooming levels anyways, in practical terms they are almost the same). After upscaling good content with NGU, I don't want to then make it blurrier with Catmull-Rom.

burfadel
1st December 2016, 09:23
I had DP to miniDP and screen always bugged on me if I left player on pause and left pc to switch screen off. Screen didn't wake up 3 out of 10 times. I got DP-DP cable it went away(for now), so I think it's driver issue in my case. It is also happening only on w10.
Just do not leave player running when you need to step away from PC.
Also it's generally bad to let PC go to sleep, it somewhat works for laptops because battery life, but there is no need to do that with PC. So many things can go wrong with modern GPU drivers, probably that's what happening, you are not saving word document into RAM, it's a bit more complicated with rendering I think.
Set it do disable display and never go to sleep, most safe option.

Yes, I have a similar problem. madVR seems to lock the player in memory, you can't end the process and you have to literally turn the power off and back on again because it locks such that you can't restart (restart just hangs).

I think I might start another thread specifically targetting any issues with madVR and the hardware involved. I will emphatically state the purpose of the thread so people don't try an argue that it works fine for them so there isn't a problem. From what I have found, the issues seem to be more prevalent on Geforce GTX 10xx models and Radeon RX models (the 400 series). The old R9-280X worked beautifully and on affriend's R9-280X (different other hardware though) runs beautifully with similar settings, so it will be interesting with the responses.

Backflash
1st December 2016, 09:27
I think I might start another thread specifically targetting any issues with madVR and the hardware involved. I will emphatically state the purpose of the thread so people don't try an argue that it works fine for them so there isn't a problem. From what I have found, the issues seem to be more prevalent on Geforce GTX 10xx models and Radeon RX models (the 400 series). The old R9-280X worked beautifully and on affriend's R9-280X (different other hardware though) runs beautifully with similar settings, so it will be interesting with the responses.

It's not madvr related. For example if you send PC to sleep/screen switch off through GPU while running a game same thing will happen.

GPU drivers must not go to sleep on load.

burfadel
1st December 2016, 09:38
It's not madvr related. For example if you send PC to sleep/screen switch off through GPU while running a game same thing will happen.

GPU drivers must not go to sleep on load.

With the R9-280X I could let the computer go to sleep if I had it paused in D3D 11 Exclusive mode, or in a windowed mode D3D 11 etc without any issues. Therefore, I know that you can let it do it, the issue and others appears to be with newer Geforce and AMD cards.

If it doesn't work for you, then you are just another person that is affected by this!

Backflash
1st December 2016, 09:43
With the R9-280X I could let the computer go to sleep if I had it paused in D3D 11 Exclusive mode, or in a windowed mode D3D 11 etc without any issues. Therefore, I know that you can let it do it, the issue and others appears to be with newer Geforce and AMD cards.

If it doesn't work for you, then you are just another person that is affected by this!

Well yeah, but it's drivers+windows 10 that does this, madvr may trigger it more often but it happens with everything else anyway.

Razoola
1st December 2016, 10:01
I wanted to give some feedback on the new settings layout. Personally I do not like it much as I feel its a little confusing. At the same time though I can see the original way was probably not the best either simply because it has elvolved over time. Both ways kind of over complicate the settings. I have a suggestion of a settings layout that may work but at the same time I guess some may not like it and it would be alot of work for madshi given it probably needs the current system demolished and rebuilt...

My idea is under 'scaling algorithms' there should be three sub option windows.

- Movie Original
- Movie Upscaling
- Movie Downscaling

Under each of these settings are dropdown boxes (much like the new doubling options) and other settings that follow the rendering process, from top of page to bottom of page. The user can then choose how chroma / doubling and everything else is handled as it passes through the rendering process. Movie original would be for example where a 1080p move is output to a 1080p disaply. The other two are self explaining.

Would something like that be workable madshi? What do you power users think? I feel it would simplify the options greatly but thats just my opinion.

burfadel
1st December 2016, 10:19
Well yeah, but it's drivers+windows 10 that does this, madvr may trigger it more often but it happens with everything else anyway.

I started a new thread here for specific issues like this:
http://forum.doom9.org/showthread.php?p=1788137

I do call it an issue because it does not affect the R9-280X with all other things equal. It seems to only affect newer cards from Nvidia and AMD. If it is driver related then tracking down the cause will allow it to be reported via driver reporting :).

ashlar42
1st December 2016, 10:34
Exactly.

Here you can see a comparison between Jinc, Lanczos3, Bicubic150 and SSim1D100 for downscaling after 2x supersampling with NGU:

https://s16.postimg.org/jm46cg02p/Jinc_AR.png (https://postimg.org/image/jm46cg02p/) https://s16.postimg.org/ytk5wsrxd/Lanczos_3_AR.png (https://postimg.org/image/ytk5wsrxd/) https://s16.postimg.org/wnpv2aogx/Bicubic150_AR.png (https://postimg.org/image/wnpv2aogx/) https://s16.postimg.org/nwiu8155t/SSim1_D100_AR.png (https://postimg.org/image/nwiu8155t/)

And zoomed:

https://s16.postimg.org/pgqu38035/Jinc_AR.png (https://postimg.org/image/pgqu38035/) https://s16.postimg.org/583ghi0s1/Lanczos_3_AR.png (https://postimg.org/image/583ghi0s1/) https://s16.postimg.org/q2fszbv5d/Bicubic150_AR.png (https://postimg.org/image/q2fszbv5d/) https://s16.postimg.org/4haqbpyep/SSim1_D100_AR.png (https://postimg.org/image/4haqbpyep/)

Make your choice. For me it is Lanczos3 because Bicubic150 has too much aliasing and Jinc is a little bit too soft.I don't find Bicubic150 to be aliased at 100% zoom, I just find it pleasantly sharper.

Ver Greeneyes
1st December 2016, 11:48
Looking at Neo-XP's zoomed in comparison, there are some spots where SSIM seems to preserve the structure of the upscale better, like a spot on the edge of the woman's face between her eyebrow and eye (on the right side). Other than that it actually seems slightly softer than Bicubic150, but I think the sharpness overall is somewhat more consistent. So I do think this proves that SSIM can make a visible difference, even if the number of pixel differences for a given image are small. Whether the trade off between letting people use all the power of their GPUs and keeping them from wasting power on inconsequential things is worth it is another question.

Backflash
1st December 2016, 12:17
Looking at Neo-XP's zoomed in comparison, there are some spots where SSIM seems to preserve the structure of the upscale better, like a spot on the edge of the woman's face between her eyebrow and eye (on the right side). Other than that it actually seems slightly softer than Bicubic150, but I think the sharpness overall is somewhat more consistent. So I do think this proves that SSIM can make a visible difference, even if the number of pixel differences for a given image are small. Whether the trade off between letting people use all the power of their GPUs and keeping them from wasting power on inconsequential things is worth it is another question.

There is a weird dot on the left eye(if looking at picture) center with SSIM though. just a glitch?

Neo-XP
1st December 2016, 12:17
I don't find Bicubic150 to be aliased at 100% zoom, I just find it pleasantly sharper.

I could live with Bicubic150 for downscaling, it just accentuate the sharpening of Lanczos3 (but the dark halos are increased too). Lanczos3 is just sharp enough for me.

I will test all the algorithms with a bad source material. It should be easier too see all the qualities and defaults of each one.

Ver Greeneyes
1st December 2016, 13:01
There is a weird dot on the left eye(if looking at picture) center with SSIM though. just a glitch?
Do you mean around the bottom right corner of the highlight (the window reflection) in her pupil? Yeah that looks a little weird, might be there in the upscale too though.

clsid
1st December 2016, 15:33
If the PC does go to standby and return the video is always green and I have to restart the player for it to work again.

With mpc-hc using the EVR it doesn't happen.
I think this is probably something that madshi can fix.

EVR-CP has similar issues after sleep/hibernate, which can be solved by turning on this option:
options > playback > output > reinitialize when changing display

Relevant code:
https://github.com/mpc-hc/mpc-hc/blob/develop/src/filters/renderer/VideoRenderers/DX9AllocatorPresenter.cpp (starts at line 1627)

clsid
1st December 2016, 16:19
Hi madshi, since you are now also working on simplifying the settings GUI, I have some ideas regarding the presets/profiles.

1) Get rid of the profile groups. It is just unnecessarily complex. Just let presets cover the whole group in each section.
2) On the main level in the GUI (e.g. scaling algorithms page) show a list of all presets in that section. Have buttons to remove and duplicate a preset.
3) Always have at least one preset. For example called "default". This one can't be removed. It is used as a fallback.
4) Display all settings pages just once. Have a dropdown box somewhere at the top to select the "current" preset. Simpler GUI and one can quickly compare settings of different presets. Have an entry called "new" at the bottom of the list. When selected, this takes the user to the preset creation page of that section.
5) Don't use a profile selection script, but instead give each preset a boolean statement that determines if it should get loaded. This has two benefits. First of all, presets become fully independent, making them easy to share/exchange. So we can start making "Doom9 community approved (TM)" optimized presets in the near future. Secondly, it is much easier to have a large number of presets. Non-programmers are often already struggling with relatively simple profile selection scripts. Given the increasing popularity of 4k and 60fps video, plus the availability of HDR and 3D functionality in madvr, the amount of presents can be significant in case of the true enthusiasts.
6) Add buttons to import/export a preset. Preferably in a (human readable) text-based format (xml/ini) for easy sharing.
7) Add a right-click menu to the boolean statement field of the preset. This is used to quickly insert commonly used variables (srcWidth, deintFps, etc) and could even have submenus for stuff like "(deintFps > 31)".

Uoppi
1st December 2016, 16:22
Otherwise, if you can run it, I find NGU high (the new high, old medium) is a breakpoint where the image becomes noticeably better.

Do you mean for luma? Or also for chroma? I tend not want to "waste" resources on chroma by going all-out without visible benefit.

I like NGU a lot but not so much for SD -> 1080p. The picture is too "frizzy" or oversharp. I'll need to play with soften edges, I guess.

Ver Greeneyes
1st December 2016, 16:34
5) Don't use a profile selection script, but instead give each preset a boolean statement that determines if it should get loaded. This has two benefits. First of all, presets become fully independent, making them easy to share/exchange. So we can start making "Doom9 community approved (TM)" optimized presets in the near future. Secondly, it is much easier to have a large number of presets. Non-programmers are often already struggling with relatively simple profile selection scripts. Given the increasing popularity of 4k and 60fps video, plus the availability of HDR and 3D functionality in madvr, the amount of presents can be significant in case of the true enthusiasts.
I'm not opposed to this, but it does mean the order in which the profiles are listed becomes significant - the first profile to match a rule would be loaded. Right now the profile selection script determines this ordering; that would need to be replaced with something less generic (like a configurable box listing the profiles in the order they will be tried).

clsid
1st December 2016, 16:47
Yes, of course, the order of the list of presets should preferably be customizable. That is easy to implement (with up/down button for moving a selected preset).

But ideally the loading rule of a preset is specific enough to avoid multiple matches. At least when we start creating presets for others.

The idea is to eventually have a few presets for each common type of situation (e.g. 720p23 -> 1080p23) with increasing quality/complexity. Plus a fixed set of sample files for testing. Plus a dedicated topic for testing results. So everyone can compare rendertimes and pick the most suitable preset for their hardware. Ultimately a little tool that automatically sets proper presets. Unless madshi beats us with some kind of automated performance based magic.

nevcairiel
1st December 2016, 16:57
As a programmer, I like the profile selection scripts. I have one single view where I can see which rules determine which profiles are used. Having to check every single profile and only seeing one line at a time would be quite a step back, imho.
As a compromise it could offer an "advanced" and "simple" mode, where in simple mode it just constructs the script from the individual lines, and in advanced mode, let me write it.

clsid
1st December 2016, 17:09
Yes, that would be a reasonable compromise.

Perhaps additionally (at a later stage) a preset "test" function could be added? A test button that open a dialog where you can specify the various parameters such as srcwidth/srcheight (textbox), fps (dropdown), filename, HDR (checkbox), etc. That then shows which preset(s) matches the chosen values.

JarrettH
1st December 2016, 17:10
How do you even tell what is what at 800% zoom?

Warner306
1st December 2016, 18:01
I don't find Bicubic150 to be aliased at 100% zoom, I just find it pleasantly sharper.

Same here. I like Bicubic150 and SSIM.

That's for downscaling. Although, the difference is so small I don't really care.

Neo-XP
1st December 2016, 19:42
Same here. I like Bicubic150 and SSIM.

That's for downscaling. Although, the difference is so small I don't really care.

Here is another test with an upscaling to UHD :

Mitchell-Netravali / Catmull-Rom / Bicubic150 / Lanczos3 / Jinc

https://s11.postimg.org/ktu2glzlr/Mitchel_Netravali.jpg (https://postimg.org/image/ktu2glzlr/) https://s11.postimg.org/5wllflmdb/Catmull_Rom.jpg (https://postimg.org/image/5wllflmdb/) https://s11.postimg.org/c7qsw0nlr/Bicubic150.jpg (https://postimg.org/image/c7qsw0nlr/) https://s11.postimg.org/pd6b24hhb/Lanczos3.jpg (https://postimg.org/image/pd6b24hhb/) https://s15.postimg.org/7syxb6y7b/Jinc.jpg (https://postimg.org/image/7syxb6y7b/)

Zoom 100% on FHD screen and lossless quality :

https://s16.postimg.org/cyj0ifaw1/Mitchel_Netravali.png (https://postimg.org/image/cyj0ifaw1/) https://s16.postimg.org/fc0y9fpb5/Catmull_Rom.png (https://postimg.org/image/fc0y9fpb5/) https://s16.postimg.org/3yeey8esh/Bicubic150.png (https://postimg.org/image/3yeey8esh/) https://s16.postimg.org/mhspbvye9/Lanczos3.png (https://postimg.org/image/mhspbvye9/) https://s16.postimg.org/xsvd0959d/Jinc.png (https://postimg.org/image/xsvd0959d/)

And the source (saved by MPC) :

https://s16.postimg.org/5g6pvfg75/source.png (https://postimg.org/image/5g6pvfg75/)

Once again, Lanczos3 wins for me.
You can easily see there that Bicubic150 produce a lot of aliasing and other ugly artifacts.
That is why I use Lanczos3 after luma doubling with NGU and why it should be available in madVR's new versions, for upscaling and downscaling.

pose
1st December 2016, 20:04
That is why I use Lanczos3 after luma doubling with NGU and why it should be available in madVR's new versions, for upscaling and downscaling.
Literally no logic there.

Neo-XP
1st December 2016, 20:19
Literally no logic there.

Maybe I wasn't explicit enough. I used a bad source material to highlight the defaults of each algorithm. When used for luma doubling, NGU will produce some artifacts that will be increased by the algorithm that you put after it, for upscaling and downscaling. Jinc is also very good if you want a clean result, but the rendering times make it not relevant for me.

har3inger
1st December 2016, 20:50
Here is another test with an upscaling to UHD :

Mitchell-Netravali / Catmull-Rom / Bicubic150 / Lanczos3 / Jinc

https://s11.postimg.org/ktu2glzlr/Mitchel_Netravali.jpg (https://postimg.org/image/ktu2glzlr/) https://s11.postimg.org/5wllflmdb/Catmull_Rom.jpg (https://postimg.org/image/5wllflmdb/) https://s11.postimg.org/c7qsw0nlr/Bicubic150.jpg (https://postimg.org/image/c7qsw0nlr/) https://s11.postimg.org/pd6b24hhb/Lanczos3.jpg (https://postimg.org/image/pd6b24hhb/) https://s15.postimg.org/7syxb6y7b/Jinc.jpg (https://postimg.org/image/7syxb6y7b/)

Zoom 100% on FHD screen and lossless quality :

https://s16.postimg.org/cyj0ifaw1/Mitchel_Netravali.png (https://postimg.org/image/cyj0ifaw1/) https://s16.postimg.org/fc0y9fpb5/Catmull_Rom.png (https://postimg.org/image/fc0y9fpb5/) https://s16.postimg.org/3yeey8esh/Bicubic150.png (https://postimg.org/image/3yeey8esh/) https://s16.postimg.org/mhspbvye9/Lanczos3.png (https://postimg.org/image/mhspbvye9/) https://s16.postimg.org/xsvd0959d/Jinc.png (https://postimg.org/image/xsvd0959d/)

Once again, Lanczos3 wins for me.
You can easily see there that Bicubic150 produce a lot of aliasing and other ugly artifacts.
That is why I use Lanczos3 after luma doubling with NGU and why it should be available in madVR's new versions, for upscaling and downscaling.

You need to show us the source. Something is definitely weird here, as the aliasing "block" size is much larger than a single pixel, if you compare to native resolution in the OSD. It's like the source is nearest-neighbor doubled/tripled or something.

Do you mean for luma? Or also for chroma? I tend not want to "waste" resources on chroma by going all-out without visible benefit.

I like NGU a lot but not so much for SD -> 1080p. The picture is too "frizzy" or oversharp. I'll need to play with soften edges, I guess.

Sorry, should have specified. I meant luma. I don't use NGU for chroma, because it's almost invariably filled with compression artifacts which the algorithm thinks is real detail a lot of the time. For very clean sources, it's quite nice, but similar results can be achieved for faster using Recon Soft or bilateral.

NGU is best for source content 720p and up scaled to either 1080p or 1440p. I don't have 4k and cannot comment. For SD and lower, nnedi3 is still king (with SR off if source is dirty enough), and could be improved with Add Grain if madshi decides to allow the combination.

kalston
1st December 2016, 21:04
Here is another test with an upscaling to UHD :

Mitchell-Netravali / Catmull-Rom / Bicubic150 / Lanczos3 / Jinc

https://s11.postimg.org/ktu2glzlr/Mitchel_Netravali.jpg (https://postimg.org/image/ktu2glzlr/) https://s11.postimg.org/5wllflmdb/Catmull_Rom.jpg (https://postimg.org/image/5wllflmdb/) https://s11.postimg.org/c7qsw0nlr/Bicubic150.jpg (https://postimg.org/image/c7qsw0nlr/) https://s11.postimg.org/pd6b24hhb/Lanczos3.jpg (https://postimg.org/image/pd6b24hhb/) https://s15.postimg.org/7syxb6y7b/Jinc.jpg (https://postimg.org/image/7syxb6y7b/)

Zoom 100% on FHD screen and lossless quality :

https://s16.postimg.org/cyj0ifaw1/Mitchel_Netravali.png (https://postimg.org/image/cyj0ifaw1/) https://s16.postimg.org/fc0y9fpb5/Catmull_Rom.png (https://postimg.org/image/fc0y9fpb5/) https://s16.postimg.org/3yeey8esh/Bicubic150.png (https://postimg.org/image/3yeey8esh/) https://s16.postimg.org/mhspbvye9/Lanczos3.png (https://postimg.org/image/mhspbvye9/) https://s16.postimg.org/xsvd0959d/Jinc.png (https://postimg.org/image/xsvd0959d/)

Once again, Lanczos3 wins for me.
You can easily see there that Bicubic150 produce a lot of aliasing and other ugly artifacts.
That is why I use Lanczos3 after luma doubling with NGU and why it should be available in madVR's new versions, for upscaling and downscaling.

Pretty sure that bicubic150 is bugging out here - there is no way in hell it should look that bad. Dunno if it's your setup or the algo though. Maybe you can share a short sample?

Neo-XP
1st December 2016, 21:09
Pretty sure that bicubic150 is bugging out here - there is no way in hell it should look that bad. Dunno if it's your setup or the algo though. Maybe you can share a short sample?

Post updated with source. I didn't change any other settings while testing, just switched the algos and saved.

Dodgexander
1st December 2016, 21:31
I'd start with the system BIOS first. Reset it and test then look for an updated BIOS for the motherboard and GPU.
Thanks, I did a bit more troubleshooting and found it seems to go away if I set my memory to its lowest timing. I am using 4x2gb sticks which I think must be straining my northbridge (Hypertransport?? since I have an AMD board) so I am going to continue running with a low frequency on the ram and see how it goes. I may need to raise the northbridge clock and/or frequency to get stable at the rams normal timings. Its so strange because I ran memtest86 for 48 hours and prime and it found no faults. Usually it catches faults between the CPU and Memory but this time it doesn't. At least I have a way to test it that works 70% of the time! thanks very much for your help.

Edit* spoke to soon. I tested going in to standby manually several times and thought it was fixed, however again when I left my computer with video paused in mpc hc and madvr it bsod and rebooted with ntoskrnl.exe VIDEO_SCHEDULER_INTERNAL_ERROR

All of my BSODs seem to be related to 3 things causing it
watchdog.sys
nviddmkm.sys
ntoskrnl.exe

all are VIDEO_SCHEDULER_INTERNAL_ERROR

i'm starting to wonder if it is driver related.

I had DP to miniDP and screen always bugged on me if I left player on pause and left pc to switch screen off. Screen didn't wake up 3 out of 10 times. I got DP-DP cable it went away(for now), so I think it's driver issue in my case. It is also happening only on w10.
Just do not leave player running when you need to step away from PC.
Also it's generally bad to let PC go to sleep, it somewhat works for laptops because battery life, but there is no need to do that with PC. So many things can go wrong with modern GPU drivers, probably that's what happening, you are not saving word document into RAM, it's a bit more complicated with rendering I think.
Set it do disable display and never go to sleep, most safe option.
Leaving it on all the time is something I debated but I am on and off my PC all day so it would save a considerable amount of power in the long term if the PC would just turn itself into standby. At the moment that is what I keep having to do, just keep forgetting not to leave it playing.

I'm not sure in my case it is a driver related problem though because reducing my memory frequency seems to have fixed it (so far)

turns out despite testing it doesn't help if I leave my computer. Seemed to work ok going in and out of standby manually though.

I started a new thread here for specific issues like this:
http://forum.doom9.org/showthread.php?p=1788137

I do call it an issue because it does not affect the R9-280X with all other things equal. It seems to only affect newer cards from Nvidia and AMD. If it is driver related then tracking down the cause will allow it to be reported via driver reporting :).

In my case I use a geforce 660 and it happens, i'm not sure its specific to recent nvidia cards anyway. I will post in your thread.

I think this is probably something that madshi can fix.

EVR-CP has similar issues after sleep/hibernate, which can be solved by turning on this option:
options > playback > output > reinitialize when changing display

Relevant code:
https://github.com/mpc-hc/mpc-hc/blob/develop/src/filters/renderer/VideoRenderers/DX9AllocatorPresenter.cpp (starts at line 1627)

Thanks, that would be awesome!

har3inger
1st December 2016, 21:37
Your source is very obviously not larger than 1440p. You shouldn't compare downscales with it, unless you're using nnedi3 or ngu.

Edit: Didn't catch this at first, but you're using the algorithms to *upscale* according to the OSD. Bicubic 150 shouldn't be used to upscale, ever. It probably shouldn't even be selectable in madvr. Of course bicubic 150 looks like garbage on upscale.

QBhd
2nd December 2016, 00:42
^^

Agreed, terrible method used for the given situation... That comparison makes zero sense

QB

glass
2nd December 2016, 03:45
Hi, I'm still getting a bug on the latest few versions where screenshots taken in MPC-HC just result in a white garbled mess. The only workaround is to scale the video first.

I'm using a GTX 1060, but it's been constant among driver versions also.

jerryleungwh
2nd December 2016, 06:53
May I ask whether it will produce a better quality image if I use

-NGU-high chroma for upscaling to produce a 1080p image then use an AV-receiver to upscale it to 4K, or

-use Catmull-Rom for chroma upscaling and NGU-low for luma doubling to upscale a 1080p image to 4K?

I do not have a 4K monitor at the moment but I plan to get one in the near future but my GPU isn't quite powerful enough to use anything higher than NGU-low for luma doubling to upscale a 1080p image to 4K, so which is a better option?

BluesFanUK
2nd December 2016, 15:43
I've got a 1440p monitor and i'm struggling to tell the difference between any of the settings now (and no I don't need glasses).

Is this only beneficial to Anime or something?

Warner306
2nd December 2016, 18:23
For those interested, Kodi DSPlayer has been updated to the latest version of Kodi (v17) and is now compatible with the new settings dialogs.

This is a surprise release. I'm not sure if it will continue to have support in the future.

omarank
2nd December 2016, 18:25
Feedback on NGU Low:

NGU low is great in itself. I actually like its lesser sharpness as compared to the higher versions of NGU. I tend to get distracted by the disproportion that NGU high and NGU very high create in the image in terms of definition of the edges and the texture. Even “soften edges” doesn’t help much there. NGU low is better in that regard. The performance is great too.

TheElix
2nd December 2016, 18:57
I'm sorry for bothering everyone with my petty problem, but what could be the cause of madVR not detecting any display device at all? I've noticed that madVR stopped switching frequencies, then I've noticed that the display that I was using isn't marked as active (bold) in madVR. Then I deleted all devices from the list. Tries resetting to defaults, installing/uninstalling madVR, checked my computer for any unnecessary codecs to no avail. The list is now still empty under "Devices" in madVR.

Log: https: //yadi .sk/d/6J1PmbAN3235Ek (sorry for spaces, won't let me post this otherwise)
I know it's completely off-topic, but I can't connect BlueSkyFRC filter with the playback chain. What's the standard troubleshooting procedure in that case?

AngelGraves13
2nd December 2016, 23:27
It's a shame that NGU isn't scalable and just doubles, requiring huge resources to downscale back to native resolution. If that could be changed, it would actually be useable for all content and not just content that needs to be doubled.

1080p->1440p uses a ton of GPU resources, so much so that I've decided against using it as it causes my GPU to get rather hot and the fans are just too loud to enjoy the movie.

Guess I'll stick with ReconSoft for Chroma and Jinc for all my scaling needs until something better comes along.

aufkrawall
3rd December 2016, 00:52
Reconstruction for chroma is quite expensive too and has little benefit over jinc, imho.

madshi
3rd December 2016, 01:06
I've just tried to reply to all the recent posts, but I'm getting server error 500 if I try to quote anybody. So the replies will have to wait until the forum is fixed again.

optimist
3rd December 2016, 01:15
@Warner306 Wow this is really good news even if aracnoz doesn't continue to support it. Maybe he will release an update for major Kodi release.

On topic: NGU is working nicely with my Nvidia 1070. Thank you madshi for this awesome algo. I have a Panasonic plasma TV and NGU doesn't seem overly sharp with very high setting. I mostly watch HD movies and tv series, not much anime. Like others I have some low quality SD content and I have to use softcubic for them to hide the compression artifact, NGU is not suitable for them. Some thing to mask the compression artifact would be much appreciated.

seiyafan
3rd December 2016, 01:38
I can tell the difference from upscaling 1080 on 1440 monitor if I am sitting 15 inches from the screen, that's about it for me.

Alamei
3rd December 2016, 01:40
This might be a silly question, but how are you all getting that rendering time step-by-step breakdown to show up in the stats? Is it a recent feature that was added?

sneaker_ger
3rd December 2016, 01:47
Create an empty file "ShowRenderSteps" in the madvr folder.

Alamei
3rd December 2016, 01:53
Excellent. Thank you for the quick response. I can normally just get the answers by searching the forum, but it currently seems to be glitching out and throwing 500 errors on half of the pages I look at (in addition to being upset with quote attempts, as madshi mentioned).

Sp00kyFox
3rd December 2016, 04:29
is the deringing filter of madvr under "processing/artifact removal" a custom filter or is there an avisynth implementation of it? I'd like to use it in a script of mine.

Asmodian
3rd December 2016, 06:59
It is a custom filter.